Systems and methods for detecting high impact non-informative features

1 . A computer-implemented method comprising:

obtaining a first dataset including a first set of features and a first set of tags;

determining a second dataset including a second set of features and the first set of tags, wherein the second set of features being transformed from the first set of features, and wherein each of the second set of features of the second dataset comprises a binary value;

predicting, by a machine learning model, a third dataset including a third set of features and a second set of tags based on the second dataset;

comparing the third dataset to the second dataset to determine a correlation between the third dataset and the second dataset, the correlation being determined based on a contribution of each feature of the third set of features according to a score representative of a similarity between the first set of tags of the second dataset and the second set of tags of the third dataset relative to a threshold;

based on the comparison, filtering at least one feature having one or more missing feature values from the first set of features of the first dataset; and

determining a training dataset corresponding to a remaining features of the first set of features of the first dataset,

wherein the machine learning model being trained using the training dataset enables improved classification accuracy by the machine learning model.

2 . The computer-implemented method of claim 1 , further comprising:

extracting the first set of tags from the first dataset; and

associating the first set of tags with the second dataset.

3 . The computer-implemented method of claim 1 , wherein determining the second dataset including the second set of features and the first set of tags further comprises:

associating a first value with each feature of the second set of features of the second dataset based on a respective value of each feature of the first set of features of the first dataset corresponding to a null value; and

associating a second value with each feature of the second set of features of the second dataset based on a respective value of each feature of the first set of features of the first dataset corresponding to a non-null value.

4 . The computer-implemented method of claim 3 , wherein the first value is a value of 0 and the second value is a value of 1.

5 . The computer-implemented method of claim 1 , wherein comparing the third dataset to the second dataset further comprises:

comparing the first set of tags of the second dataset to the second set of tags of the third dataset;

determining the score representative of the similarity between the first set of tags and the second set of tags;

determining the contribution of each feature of the third set of features in response to the score exceeding the threshold; and

identifying the at least one feature having the one or more missing feature values based on the determined contribution.

6 . The computer-implemented method of claim 5 , wherein determining the contribution of each feature of the third set of features further comprises:

associating a value to each feature of the second set of features of the second dataset;

comparing each feature of the second set of features of the second dataset to an average to determine an impact each feature of the second set of features of the second dataset has in the prediction of the third dataset by the machine learning model; and

ranking the third set of features of the third dataset based on the determined contribution,

wherein the at least one feature having the one or more missing feature values is identified based on the ranking of the third set of features.

7 . The computer-implemented method of claim 1 , further comprising:

training the machine learning model using the second set of features of the second dataset, the machine learning model being configured to predict the third set of features and the second set of tags of the third dataset based on the second set of features of the second dataset; and

training the machine learning model using the training dataset.

8 . The computer-implemented method of claim 7 , wherein the trained machine learning model is configured to be utilized across one or more architecture domains.

9 . The computer-implemented method of claim 1 , wherein the at least one feature having the one or more missing feature values corresponds to a non-informative feature having a positive impact on the prediction by the machine learning model.

10 . A system comprising:

a processor; and

a non-transitory computer readable media having stored thereon instructions that are executable by the system to cause the processor to perform one or more operations comprising:

obtain a first dataset including a first set of features and a first set of tags;

extract the first set of tags from the first dataset;

determine a second dataset including a second set of features, wherein the second set of features is transformed from the first set of features, and wherein each of the second set of features of the second dataset comprises a binary value;

associate the first set of tags with the second dataset;

predict, by a machine learning model, a third dataset including a third set of features and a second set of tags based on the second dataset;

compare the third dataset to the second dataset to determine a correlation between the third dataset and the second dataset;

based on the comparison, filter at least one feature having one or more missing feature values from the first set of features of the first dataset; and

determine a training dataset corresponding to a remaining features of the first set of features of the first dataset,

wherein the machine learning model being trained using the training dataset enables improved classification accuracy by the machine learning model,

wherein comparing the third dataset to the second dataset to determine the correlation between the third dataset and the second dataset comprises:

determining a contribution of each feature of the third set of features in response to a score representative of a similarity between the first set of tags of the second dataset and the second set of tags of the third dataset exceeding a threshold; and

wherein the at least one feature having the one or more missing feature values being identified based on the determined contribution.

11 . The system of claim 10 , wherein determining the second dataset including the second set of features and the first set of tags further comprises:

associating a first value with each feature of the second set of features of the second dataset based on a respective value of each feature of the first set of features of the first dataset corresponding to a null value; and

associating a second value with each feature of the second set of features of the second dataset based on the respective value of each feature of the first set of features of the first dataset corresponding to a non-null value.

12 . The system of claim 11 , wherein the first value is a value of 0 and the second value is a value of 1.

13 . The system of claim 10 , wherein comparing the third dataset to the second dataset further comprises:

comparing the first set of tags of the second dataset to the second set of tags of the third dataset; and

determining the score representative of the similarity between the first set of tags and the second set of tags; and

identifying the at least one feature having the one or more missing feature values based on the determined contribution.

14 . The system of claim 13 , wherein determining the contribution of each feature of the third set of features further comprises:

associating a value to each feature of the second set of features of the second dataset;

comparing each feature of the second set of features of the second dataset to an average to determine an impact each feature of the second set of features of the second dataset has in affecting the prediction of the third dataset by the machine learning model; and

ranking the third set of features of the third dataset based on the determined contribution,

wherein the at least one feature is identified based on the ranking of the third set of features.

15 . The system of claim 10 , further comprising:

training the machine learning model using the second set of features of the second dataset; and

training the machine learning model using the training dataset,

wherein the machine learning model being configured to predict the third set of features and the second set of tags of the third dataset based on the second set of features of the second dataset, and wherein the trained machine learning model is configured to be utilized across one or more architecture domains.

16 . The system of claim 10 , wherein the at least one feature having the one or more missing feature values corresponds to a non-informative feature having a positive impact on the prediction by the machine learning model.

17 . A computer-program product comprising a non-transitory computer-readable media having stored thereon instructions that are executable by a computing system to cause a processor to perform operations comprising:

obtain a first dataset including a first set of features and a first set of tags;

determine a second dataset including a second set of features and the first set of tags, wherein the second set of features is transformed from the first set of features, and wherein each of the second set of features of the second dataset comprises a binary value, the determining of the second dataset comprising:

associate a first value with each feature of the second set of features of the second dataset based on a respective value of each feature of the first set of features of the first dataset corresponding to a null value,

associate a second value with each feature of the second set of features of the second dataset based on the respective value of each feature of the first set of features of the first dataset corresponding to a non-null value,

extract the first set of tags from the first dataset, and

associate the first set of tags with the second dataset;

train a machine learning model using the second set of features of the second dataset;

predict, by the machine learning model, a third dataset including a third set of features and a second set of tags based on the second dataset;

compare the third dataset to the second dataset to determine a correlation therebetween;

based on the comparison, filter at least one feature having one or more missing values from the first set of features of the first dataset;

determine a training dataset corresponding to a remaining features of the first set of features of the first dataset; and

training the machine learning model using the training dataset,

wherein the machine learning model being trained using the training dataset enables improved classification accuracy by the machine learning model,

wherein comparing the third dataset to the second dataset to determine the correlation therebetween comprises:

determining a score representative of a similarity between the first set of tags and the second set of tags;

determining a contribution of each feature of the third set of features in response to the score exceeding a threshold,

wherein the at least one feature is filtered from the first dataset based on the determined contribution.

18 . The computer-program product of claim 17 , wherein the first value is a value of 0 and the second value is a value of 1.

19 . The computer-program product of claim 17 , wherein comparing the third dataset to the second dataset further comprises:

comparing the first set of tags of the second dataset to the second set of tags of the third dataset; and

identifying the at least one feature based on the determined contribution.

20 . The computer-program product of claim 19 , wherein determining the contribution of each feature of the third set of features further comprises:

associating a value to each feature of the second set of features of the second dataset;

comparing each feature of the second set of features of the second dataset to an average to determine an impact each feature of the second set of features of the second dataset has in affecting the prediction of the third dataset by the machine learning model; and

ranking the third set of features of the third dataset based on the determined contribution,

wherein the at least one feature having the one or more missing values is identified based on the ranking of the third set of features,

wherein the at least one feature corresponds to a non-informative feature having a positive impact on the prediction by the machine learning model.